Every leader operates with a margin. It is the room to absorb a bad quarter, a hostile headline, a supply shock, an election that goes the wrong way. For most of the past two decades, that margin was wide. Leadership in politics and in business could afford to look confident without being precise.
That margin has narrowed sharply. The narrowing is structural, not a passing cycle. Two industries, on the surface unrelated, show the shift with unusual clarity.
The first is shipping. A vessel routing through the Red Sea today carries a cost structure that would have been unrecognisable three years ago. Insurers price war-risk premiums by the transit, sometimes by the day. Shipowners weigh two costs against each other. One is rerouting around the Cape of Good Hope: weeks of extra fuel, extra time, extra capital tied up in transit. The other is proceeding through waters where naval escort has become a precondition, not a courtesy.
Hormuz sits behind all of this as the deeper structural risk. A fifth of the world’s oil moves through this one chokepoint. It is governed less by freedom-of-navigation guarantees than by a working assumption: no one wants to be the party that closes it. Shipping executives now build annual plans around a variable that used to be someone else’s problem. The credibility of rule-of-law enforcement in international waters has become a line item.
The second is artificial intelligence. For a decade, the constraint on AI economics was compute: chips, capital, talent. Boards and investors modelled the industry accordingly. That constraint has not disappeared. It has been overtaken by a slower, heavier one: electricity. Training and running frontier models at scale now runs into physical limits: grid capacity, transmission buildout, the multi-year timelines of power generation projects. A data centre can be financed and built faster than the substation that feeds it. Companies with the strongest AI ambitions are discovering where their real planning horizon sits: the next decade of energy infrastructure, set by utilities, regulators, unions, local communities and permitting authorities they do not control.
What connects a shipping line and an AI lab is the position both have been put in. Both are designing strategy around a constraint that persists rather than passes, one that sits largely outside their own authority to fix.
Academic research has been circling this problem from two directions that rarely meet.
Management scholarship has a name for coping with limits well: bounded leadership. The concept, developed by Andrzej Koźmiński, treats a leader’s effectiveness as shaped by competence and by something else too: how skilfully that leader navigates a specific set of constraints, among them power dynamics, organisational culture, rules, and access to information. A related body of work reviewed well over a hundred empirical studies on creativity and innovation. It found that constraint sharpens performance up to a threshold, and only degrades it once it becomes extreme.
Limits, handled deliberately, can be a source of strength.
Political science approaches the same territory with a different emphasis. Research on political decision-making identifies a leader’s truly scarce resource as attention itself. It is the capacity to prioritise among competing demands within a fixed cognitive budget, an idea tracing back to Herbert Simon’s work on bounded rationality. A separate strand measures political constraint as a structural feature of a system, set by constitutions, coalitions, and electoral rules. It holds largely independent of who happens to be leading at a given moment.
The difference in emphasis is instructive. Business scholarship tends to treat constraint as material a leader can work with, and occasionally convert into advantage. Political science tends to treat it as a condition a leader operates inside, set by institutions and events that precede any individual’s arrival in office.
Watching heads of state and boards manage the current environment, both descriptions apply at once. That is precisely what makes the moment demanding. A head of state confronting a chokepoint like Hormuz has one lever available: candour. No amount of posturing will move an international-law reality, so the only credible move is naming the constraint plainly rather than pretending it can be negotiated away. A board confronting the same chokepoint has a different lever. It is not answerable for the constraint’s existence, only for how well it manoeuvres inside it: rerouting, insuring, diversifying.
Political leadership carries the burden of legitimising a constraint it did not choose, and that burden is discharged through honesty rather than argument. Business leadership carries the burden of exploiting whatever discretion survives inside a constraint it did not choose either, and that burden is discharged through movement rather than explanation. The discipline required in each case differs in degree, though it draws on the same instinct.
This is where leadership is being tested differently than it was even five years ago, in politics and in business alike. A head of state managing energy security, coalition arithmetic, or sanctions exposure faces the same underlying discipline as a board managing capital allocation under geopolitical risk or power-constrained growth. In both cases, the room to manoeuvre has narrowed, and it is not coming back on its own timeline. The leaders who look credible right now have absorbed that fact into how they plan, rather than treating it as an obstacle between them and a return to how things were.
Absorbing it looks like a specific set of behaviours. They are consistent whether the leader sits in a cabinet room or a boardroom.
The first is sequencing: focusing on the constraint that actually binds, not the one that is easiest to discuss publicly. That means energy capacity rather than chip supply. It means insurance and transit risk rather than freight rates alone.
The second is optionality. The shipping companies managing this best have kept both routes live. They have not committed fully to either the Red Sea staying open or the Cape becoming the new default. The AI companies managing it best are securing power commitments years ahead of the demand curve. They are building the constraint into their planning horizon instead of discovering it later.
The third applies to government: naming the actual limit to the electorate or the market, rather than the more comfortable version of it. The credibility cost of being caught wrong about your own constraint is far higher, and far more lasting, than the cost of stating it plainly.
What this requires, above all, is precision: an accurate account of where the real room is, communicated without spin, and spent with intent rather than urgency.
A board that treats a capital or energy constraint as temporary will keep making decisions sized for a world that no longer exists. A government that treats a security or fiscal constraint the same way will do the same. The cost shows up later. It shows up as a credibility gap, with markets, with citizens, with counterparties. That gap is much harder to close than the original constraint would have been to plan around.
Shipping and AI simply make the pattern visible in numbers: insurance premiums, gigawatts, transit days. Most political and business leadership is now living with some version of the same condition. Margin is narrower. The constraints that actually matter now play out over longer horizons. Patience is thinner from the people they answer to, for leaders who have not noticed the shift.
Management research treats this as material to work with. Political science treats it as a structural condition to be honestly reckoned with. Both are right. The leaders worth watching in the years ahead have already built their plans on the assumption that the room stays this narrow. The rest are still waiting for it to widen.
